Every year, millions of Americans present at the hospital with knee injuries, such as meniscus tears or anterior cruciate ligament (AGL) sprains. Moreover, knee injuries are one of the most common causes of missed workdays. The current paradigm of treating knee injuries initially involves frequent physical therapy visits, subjective evaluations by experts, and possibly surgery; following these initial steps the patient continues to participate in physical therapy, periodically - and typically infrequently - returns to the clinic for follow-up subjective evaluations, and bases his I her joint health rehabilitation status mainly on symptoms and pain. There is no technology available currently to provide patients with knee injuries frequent, objective, and in-depth information regarding the status of their joint rehabilitation. The hypothesis for this project is that the sounds of the joints measured using sensors embedded in a wearable wrap can provide a clinically-relevant biomarker for joint health rehabilitation assessment, and can ultimately allow patients to tune their rehabilitation exercises dynamically based on objective feedback. This could potentially accelerate rehabilitation, reduce the risk of re-injury, and empower patients to be in control of their rehabilitation. This project proposes to study these sounds, and their measurement, with an integrative program including the following specific aims: (1) Design and implement an ultra-low noise, high-bandwidth, wafer-level-packaged micro-accelerometer chip for contact measurement of joint sounds from the skin surface with high fidelity; (2) Elucidate the origin of the sounds and how they change with injury using a cadaver model; (3) Develop algorithms for extracting salient features from the joint sound signals that can be used to assess joint health; (4) Evaluate the sensors and analytics in a population of 20 subjects with meniscus tears, before and after surgery, and twice during rehabilitation several months following surgery.